Providing field-programmable devices in a distributed execution environment

1. A computer-implemented method for providing field-programmable gate arrays (“FPGAs”) in a distributed execution environment, the computer-implemented method comprising: receiving; from a customer computing system, a request to launch a virtual machine instance, the virtual machine instance providing access to field-programmable device resources within the distributed execution environment, the request to launch the virtual machine instance specifying a first configuration for the field-programmable device resources, the first configuration including a first type of field-programmable device resources and a first number of field-programmable device resources; in response to receiving the request to launch the virtual machine instance, launching the virtual machine instance; receiving, by the virtual machine instance, an FPGA design for use in programming the field-programmable device resources; programming the field-programmable device resources based at least part on the FPGA design and the first configuration to produce a first programmed FPGA; causing the first programmed FPGA to run within the distributed execution environment; returning results of running the first programmed FPGA to the customer computing system; based at least in part on the results, performing an auto-scaling function to modify the first configuration used to program the first programmed FPGA to produce a second configuration, the second configuration including a second type of field-programmable device resources and a second number of field-programmable device resources, wherein the second type of field-programmable device resources differs from the first type of field-programmable device resources or the second number of field-programmable device resources differs from the first number of field-programmable device resources; reprogramming the field-programmable device resources based at least in part on the FPGA design and the second configuration to produce a second programmed FPGA; and causing the second programmed FPGA to run within the distributed execution environment.

2. The computer-implemented method of claim 1 , wherein the first programmed FPGA or the second programmed FPGA comprises a physical FPGA.

3. The computer-implemented method of claim 1 , wherein the first programmed FPGA or the second programmed FPGA comprises a virtual FPGA.

4. The computer-implemented method of claim 3 , wherein the virtual FPGA comprises a portion of a physical FPGA that is configured to provide a plurality of virtual FPGAs.

5. The computer-implemented method of claim 3 , wherein the virtual FPGA comprises a plurality of physical FPGAs.

6. The computer-implemented method of claim 1 , wherein the first programmed FPGA or the second programmed FPGA is one of a plurality of FPGAs to which the virtual machine instance provides access.

7. The computer-implemented method of claim 1 , wherein receiving the FPGA design for use in programming the field-programmable device resources comprises receiving the FPGA design via an FPGA compiler application executing on the virtual machine instance.

8. The computer-implemented method of claim 1 , wherein receiving the FPGA design for use in programming the field-programmable device resources comprises receiving the FPGA design from an FPGA design image storage component associated with the virtual machine instance.

9. The computer-implemented method of claim 1 , wherein receiving the FPGA design for use in programming the field-programmable device resources comprises receiving the FPGA design from the customer computing system.

10. The computer-implemented method of claim 1 , further comprising receiving a request to send the FPGA design to a fabrication plant.

11. A computer-implemented method comprising performing computer-implemented operations for: launching an instance within a distributed execution environment, the instance being configured to provide access to one or more field-programmable device resources using a first configuration including a first type of field-programmable device resources and a first number of field-programmable device resources; receiving a field-programmable device resource design configured to program the one or more field-programmable device resources; programming the one or more field-programmable device resources using the field-programmable device resource design and the first configuration to produce a first programmed device resource; causing the first programmed device resource to run in the distributed execution environment; returning output from the first programmed device resource to a customer computing system; based at least in part on the output, performing an auto-scaling function to automatically scale up or down either the first type of field-programmable device resources or the first number of field-programmable device resources to create a second configuration, the second configuration including a second type of field-programmable device resources and a second number of field-programmable device resources, wherein the second type of field-programmable device resources differs from the first type of field-programmable device resources or the second number of field-programmable device resources differs from the first number of field-programmable device resources; reprogramming the one or more field-programmable device resources using the field-programmable device resource design and the second configuration to produce a second programmed device resource; and causing the second programmed device resource to run.

12. The computer-implemented method of claim 11 , wherein the first programmed device resource or the second programmed device resource comprises one or more physical field-programmable devices.

13. The computer-implemented method of claim 11 , wherein the first programmed device resource or the second programmed device resource comprises one or more virtual field-programmable devices operating on one or more physical field-programmable devices.

14. The computer-implemented method of claim 11 , further comprising receiving, from the customer computing system, a request to launch the instance, and wherein the request comprises the field-programmable device resource design.

15. The computer-implemented method of claim 14 , wherein the instance is further configured to provide access to computing resources, and wherein the request further comprises a configuration for the computing resources.

16. The computer-implemented method of claim 15 , wherein receiving the field-programmable device resource design comprises receiving the field-programmable device design via an application executing on the computing resources of the instance.

17. The computer-implemented method of claim 11 , wherein the auto-scaling function is initiated based at least in part on one or more pre-defined conditions.

18. The computer-implemented method of claim 17 , wherein the one or more pre-defined conditions comprises a condition that the first programmed device resource fails to run.

19. The computer-implemented method of claim 11 , wherein the first programmed device resource or the second programmed device resource comprises one of a field-programmable gate array resource, a field-programmable object array resource, a memristor array resource, or a programmable neural network device.

20. A system comprising: a field-programmable device resource operating within a distributed execution environment; and a computer system comprising a processor and memory and operating within the distributed execution environment, the computer system being configured to; receive a request to instantiate a new instance to utilize one or more field-programmable device resources, the request specifying a first configuration for the one or more field-programmable device resources including a first type of field-programmable device resources and a first number of field-programmable device resources; instantiate the new instance to utilize the one or more field-programmable device resources; receive a field-programmable device resource design; program the one or more field-programmable device resources using the field-programmable device resource design and the first configuration to produce a first programmed device resource; cause the first programmed device resource to run within the distributed execution environment; return results of running the first programmed device resource to a customer computing system; based at least in part on the results, performing an auto-scaling function to automatically scale up or down either the first type of field-programmable resource or the first number of field-programmable resources used to run the first programmed device resource to create a second configuration, the second configuration including a second type of field-programmable device resources and a second number of field-programmable device resources, wherein the second type of field-programmable device resources differs from the first type of field-programmable device resources or the second number of field-programmable device resources differs from the first number of field-programmable device resources; reprogramming the one or more field-programmable device resources using the field-programmable device resource design and the second configuration to produce a second programmed device resource; and causing the second programmed device resource to run.

21. The system of claim 20 , wherein the computer system operates within a data center associated with the distributed execution environment.

22. The system of claim 21 , wherein the request to instantiate the new instance is received from the customer computing system that is in communication with the distributed execution environment via a network.

23. The system of claim 22 , wherein the new instance to utilize the one or more field-programmable device resources is instantiated for a cost to a customer associated with the customer computing system.